Column: Disturbing datasets

Feb 08 2014, 02:44 IST
Comments 0
SummaryGrowth data revisions make policy-decisions based on old data questionable

Two recent economic releases from the government are both encouraging and disturbing. The revision in GDP growth number for FY13 to 4.5% from 5% provides a lower base for FY14 which will provide the statistical impetus in an otherwise disappointing environment. The fact that our fiscal deficit is 95% of the targeted amount is positive for achieving the redline figure of 4.8% of the GDP for this year. But where, then, is the cause for concern?

Revisions in GDP data have become a habit in the last four years. There are significant divergences from the initial number released in May and the revised number a year later in January and the final number. In FY09, the CSO got the number right all through, at 6.7%. In case of FY10, the initial number of 7.4% went up to 8% and then 8.6%. For FY12, it moved from 8.5% to 8.4% to 8.9% while the FY12 number was scaled down from 6.5% to 6.2% but finally reverted to 6.5%. The issue is that with such revisions, the sanctity of the data gets questioned. In fact, policy decisions taken based on such numbers, in retrospect, would sound inappropriate.

There are two problems here. The first is that data collation is a challenge considering that ours is a complex economy where the unorganised sector dominates real estate, retail trade, business services, eateries, etc, cannot be tracked and their output would be subject to imputations. The second issue is conceptual. Ideally, one should not be changing weights used—as has been done this time, by using ASI data instead of IIP. While the CSO release says that the changes are marginal, it still does cause confusion in interpretation of the data. Ideally, any change in methodology should be done when the basic index or variable is being benchmarked with a new base year.

Carrying this problem further, we have seen substantial revisions in both, the WPI and the IIP, this year which poses a problem for RBI. When Governor Raghuram Rajan presented his first policy, he would have used the August WPI inflation figure of 6.1%—which actually is 7%—while, in the October review, he could have gotten more aggressive if he knew that the September inflation figure was not 6.5% but 7.1%. In fact, for the year so far, the standard deviation for the differences between initial and final inflation figures was 0.37 which works out

Single Page Format
Ads by Google
Reader´s Comments
| Post a Comment
Please Wait while comments are loading...